Guest post by Steve Forsyth, HP Software Database and Middleware R&D
Our team in R&D spends a lot of time exploring how we can help customers handle the complexities of implementing Database as a Service, or DBaaS for their production clouds. However, the key word here is not “cloud”—it is "production." Deploying DBaaS for a dev/test cloud is just not the same thing. In production environments, database resources will typically have reasonably long lifetimes.
In a production private cloud, lifecycle management cannot simply stop at create-run-stop-delete. Databases in production, even if they are "cloud," will require the essential lifecycle management activities that are required in data centers. Successful implementations of DBaaS for production must enable the operations team to visualize their database estate, understand compliance and apply remediation in a structured fashion.
Let’s take a look at the implications that DBaaS for production clouds can have for your IT organization.
DBaaS for a production private cloud must have the flexibility to provision HA/DR systems. Typically, this requires both physical and virtual server configurations in addition to DB specific configuration. This places one of our first requirements on DBaaS for Production: it must combine Infrastructure-as-a-Service (IaaS) resources with automation that can configure production-level high-availability disaster recovery (HA/DR) database systems.
Compliance and Remediation
But lifecycle management is critical, and compliance reporting and remediation are important aspects. In the context of production databases, compliance must encompass regulatory (e.g. PCI, SOX, etc.), internal standards, and relational database management system (RDBMS) specific patch levels.
Remediation tactics are fairly simple to implement and span password resetting to patch application workflows. So the real complexity with remediation is the need to integrate the tactics with the organization's broader change processes. The DBaaS API approach enables tight integration with any existing IT process automation runbook. In some more agile environments, processes may even enable putting the remediation tactics directly under the end subscriber's control.
Application Deployment and Management
Once DBaaS has provisioned an HA database and the operators can keep things under control, you still have to get the application using it. As with remediation, the challenges are not related to the development of runbooks to release database code, but assigning the right person to actually pull the trigger. In a traditional data center, the application team would most likely submit the request to the operations team to release the code, and DBaaS for production should continue to enable this model.
However, as customers become more agile with self-service, the application and operations teams may want to consider enabling an end user such as an Application Owner for production subscriptions to release the code themselves. Through code release automation, DBaaS keeps track of what has happened, so even as the operator is removed from the change control process, knowledge of the production environment state remains captured.
Another key feature of a production DBaaS implementation is Test Data Management (TDM), defined as the ability to extract data from production to use in QA environments. While RMAN, Data Pump, etc., enable the movement of the data, TDM processes must also enable the obfuscation of fields with sensitive data. For example, many of our financial services customers must ensure that credit card numbers are obfuscated in QA, with the complexity that the obfuscation must maintain the data integrity required for successful testing.
Dev/Ops and Hybrid Clouds
Some leading edge customers are also exploring sophisticated dev/ops paradigms that use application models (e.g. TOSCA). The DBaaS API approach enables these environments to access the same production grade features available from the service broker model. Another attribute with our leading edge customers is that they want hybrid clouds (i.e. a mixture of public and private resources). By following established plug-in API patterns, DBaaS can support these environments.
Creating a flexible structure
The DBaaS API helps simplify much of the inherent complexity of production environments, offering operations teams the flexibility to visualize their database estate, understand compliance and apply remediation in a structured fashion.
Share your thoughts
Over the next several weeks, some of the experts from the lab will go into more depth about how you can address common challenges of a production DBaaS. But we would also like to hear from you: How do you expect to use a DBaaS in production? Post a comment and let us know what questions you have!
● Watch a 2-minute video overview of how HP Database and Middleware Automation automates patching, provisioning, code release and compliance audit
● Attend the in-depth online presentation: Learn the ABCs of Hybrid Delivery. This 30-minute web event demonstrates how automation can streamline and optimize tasks such as: provisioning of database, virtual machines (VMs), servers, networks and storage; configuration updates; patch updates; and compliance reporting
● Watch a 4-minute video, “HP Cloud Automation: Deliver end-to-end automated infrastructure to application services” for a brief look at how hybrid service delivery is a process that requires automation as a prerequisite